Ernest I.

Becker

Polytechnic Institute of Brooklyn

Brooklyn 1, N e w York

Base-Catalyzed Alkylations
with Alcohols

Alkylation reactions with alcohols under

basic conditions have been known since 1892. Examples are the self-condensation of alcohols (the Guerbet
reaction), the alkylation of camphor, aniline, pyrrole,
malonic ester, certain phenols, and hydrocarbons.
Probably because the several reactions have not been
correlated their subject has not found its way into textbooks of organic chemistry, except for mention of the
Guerbet reaction. Despite this slight treatment, the
reaction has industrial importance in the synthesis of 2methyl-1-pentanol, in the alkylation of aniline with
methanol, and possibly in the synthesis of higher alcohols in the Synthiie process.
Although the reaction is one of alcohols, pedagogically the subject would be better treated as a part of
carbonyl condensation reactions. Reasons for this will
become evident under the discussion of the mechanism
of the reaction.
Consideration of the reaction briefly a t the undergraduate level or in more detail at the graduate level
can be a powerful teaching unit. By means of this
reaction one can bring together carbonyl condensation
reactions, the Meerweiu-Ponndorf-Verley reduction,
and the dipolar character of certain unsaturated hydrocarbons.
It is the purpose of this paper to collect the reactions
which are base-catalyzed alkylations with alcohols and
to indicate their scope and mechanism.'
Haller's alkylation of camphor with benzyl alcohol is
the first reported base-catalyzed alkylation with alcohols. I t resulted in the formation of a mixture of
benzylborneols (equation

The next example of this kind of reaction is that of M.

F. Carroll who showed that cinnamyl alcohol could alkylate ethyl acetoacetate or ethyl malonate (92, 3)
(equation 2). The reaction was also successful with

the secondary alcohol methyl phenyl carbinol. Receutly, Sprinzak has shown that active methyl groups
(4) may be alkylated (equation 3) and also active methThe appearance of higher alcohols in the Fischer-Tropsch
reaction (Synthine process) may also be due in part to Guerbettype condensations. The evidence is that adding alkali metal
oxides (and Ca) to a catalyst causes the higher rtlcohols to appear.
P. H., "Catdysis," Vol. 4.
For a recent discussion, see EMMETT,
Thesymbol +is used to represent the phenyl group.

ylene groups, such as in phenylacetonitrile (6) (equation

4).

m
(4)

A recent development has been the alkylation of the

active methylene group of hydrocarbons. The first
such example using an alcohol was 1,2,3,Ptetraphenyl-

m'

'm

cyclopentadiene (6, 7) which could be alkylated with

ethanol and with ethylene glycol in the presence of base
(equations 5, 6). In the latter case the product was
isolated as an octaphenylbifulvene, rather than the
expected bis(tetraphenylcyclopentadienyl)ethane.
Almost simultaneously with this work Sprinzak reported that fluorene could be alkylated with benzyl alcohol (equation 7) (8). Parenthetically, it should be
noted that preceding the above experiments, fluorene
had been benzylated upon heating it with henzyl ben-

zoate and sodium (9). Alkylation of fluorene has been

extended now to include normal aliphatic alcohols (lo),
secondary alcohols including isopropyl alcohol as well as
several cyclic alcohols (If), diols (It), and two dialkylaminoethanols (15). The reaction with diols merits
additional comment.'
8 NOTEADDED
IN PROOF:An analytioal application has been
reported for the alkylation reaction ( 1 % ~ ) . It was shown that

cH30w
OCHs OCHl

C:R=CHI

/ \

I \ /

OCHl

D: = H

compoundc ( R = CH8) is "methyl D" because it is obtained

from D (R = H) by heatingpithmethanol and sodium metbylate.
Volume 36, Number 3, Morch 1959

119

Dialkylation, as expected, mas observed with ethylene, 1, 2-propylene, tetramethylene, and pentamethylene glycols (equation 8). However, with glycerol the
same glycol was obtained as with ethylene glycol. With

trimethylene glycol, no product was obtained; and with

2,2dimethylpropan-l,3-diol, only 9-isobutylfluorene
was obtained (equation 9) indicating the loss of one carbon atom during the alkylations when the hydroxyl
CHa

+ HOCHIC-CHsOH-

CH&H(CHs)s

(9)

groups are 1,3. The ethylene glycol product obtained

from the glycerol reaction was identified by its melting
point, mixture melting point, and by its infrared spectrum which was identical with that obtained less deviously from ethylene glycol itself. 9-Isobutylfluorene
was synthesized independently from isobutyl alcohol
and was identical with that obtained from 2,2-dimethylprnpan-1,3-diol. Possible routes for the loss of the
carbon atom will be discussed under Mechanism.&
Alkylation of the active methylene groups of hydrocarbons has also been extended to indene, the only
hydrocarbon subjected to the reaction which has more
than one unsubstituted corner on the cyclopentadienyl
ring (1.4). Using benzyl alcohol, two benzyl groups or
one benzyl and one henaylidene group are substituted
(eauation 10).
'NOTE ADDEDIN PROOF: Wenkert and Bringi have just reported that W-2 Raney nickel catalyzes the conversion of isatin
3-ethylenethioketd (A) to 3-alkyloxindole (C) in refluxing alcohol.
He extended this to the alkylation of oxindole (B) itself and indicated that it was the intermediate in the former reaction. In a
letter dated November 25, 1958, Professor Wenkert has kindly

informed me that ethyl malonate is not dkylated with ethanol

under these conditions, while ethyl acetoacetate, acetylacetone,
and dihydroresorcinol are alkylated in low yield. Although this
Gdkylation is not novel, it is indeed interesting that it proceeds
under relatively mild conditions. Professor Wenkert feels there
is special significancein the mild reaction conditions (letter dated
December 11, 1958). However, fluorene has been alkylated
with isopropyl dcohol in boiling xylene (reaction temperature
130-140') (11).
I am indebted to Professor Wenkert for supplying results
E., AND BRINGI,N. V., J. Am.
prior to publication. (WENKERT,
Chem. Soe., 80, 5575-6(1958).)

120

Journal of Chemical Education

The most generally known of the base-catalyzed

reactions is the self-condensatinn of alcohols (16, 16,
17), now called the Guerbet reaction after its inventor.
In this reaction two molecules of an alcohol form a dimer
by the formal loss of water in such a way as to give a
carbon skeleton corresponding to that of the cnrresponding aldehyde condensation (equation 11). Both
primary and secondary alcohols condense successfully

and mixed condensations have also been effected. A

more extensive review of the Guerbet reaction has been
given recently by Machemer (18).
As far back as 1912 the alkylation of pyrrole with
ethanol and sodium ethoxide was reported (equation 12)
(19, 20). The reaction was extended to indole by

Cornforth and Robmson (21) (equation 13). These

authors converted indole and indole-2-carboxylic acid
COOH
H

qcHa
(13)

into skatole using sodium methylate and methanol.

They also found that the carboxylic acid could be
alkylated with primary alcohols, benzyl alcohol, rphenylpropyl alcohol under similar conditions. However. the acid could not be alkylated with isopropyl
alcohol, a secondary alcohol. Alternatively, indole was
successfully converted to 3-isnpropylindole using isopropan01 and sodium isopropylate and 7-methylmdole to
3-cyclohexyl-7-methylindole using cyclohexanol and
sodium cyclohexylate. I t has further been found that,
in pyrroles, ethoxycarbonyl or acetyl groups are lost
during the reaction and are replaced by alkyl (20, 22,
23) (equations 14, 15). The deacylation is effective

also with a nitrogen analognethe ketazine of 2,3,5trimethyl-4-acetylpyrrole (19, 24) (equation 16).

I n a recent paper Sprinzak has reported an unusual

alkylation of quinoline in the 3-position (4). It has
been further shown that a methyl or phenyl in the carbocyclic ring does not inhibit the alkylation (25)
(equation 17).

(and reduced
analogues)

The last category of compounds alkylated with alcohols and their alkoxides are phenols. Resorcinol.
phloroglucinol, and 2-naphthol are successfully alkylated with methanol and sodium methylate (26).
Resorcinol and phloroglucinol are both converted to
mesorcin (2,4-dihydroxymesitylene). The yields are
improved if the phenol is first converted to a methylenehis derivative or into an N-piperidylmethyl derivative.
In this way 2,2'-dihydroxydi-or-naphthylmethanegives
75% of 1-methyl-2-naphthol, the best procedure for
making this compound (equation 18). The corre-

sponding benzylidene derivative can also be used, in

this case giving rise to a mixture of 1-methyl-2-naphthol,
2-naphthol, and 1-benzyl-2-naphthol. l-(lf-Piperidylmethyl)-2-naphthol, 4-(1'-piperidylmethyl)-1-naphthol,
and 2,7-dihydroxy-1,s-bis(lf-piperidylmethy1)phenanthrene all give the corresponding methyl compounds.
It is noteworthy that 4-(1'-piperidylmethyl)-1-naphthol
gives 4-methyl-1-naphthol since 1-naphthol does not
undergo the reaction. The authors also determined
that phenol, catechol, and quinol as well as 1-naphthol
were recovered unchanged.
The Mechanism of the Reaction

gen, but only one, also fail t o be alkylated: 9-alkylfluorenes, triphenylmethane (28), monoalkylated ethyl
malonate and ethyl acetoacetate. More than one
alkyl group can be introduced if the compound may
rearrange t o give a new potential methylene group.
Thus, 3-ethyl-4-methylpyrrole gives 2,3diethyl-4methylpyrrole and 2,3,5-triethyl-4-methylpyrrole(to), but
no more than one alkyl group is introduced per carbon
atom in the ring and no alkyl group is introduced on the
nitrogen. It has also been found that 1,2,3,4-tetraphenylfulvene, a related molecule, is reduced with
sodium ethylate or butylate under the same conditions
to 5-methyl-1,2,3,4tetraphenylcyclopentadie(SO).
A base is required for the reaction. Guerbet's
early work established this point. This may be questioned with amines, since added base is not a1wa.y~employed (Sf); however, the amine may serve as its own
base. I n the alkylation of fluorene, need for a base has
been established (Id).
Recently, the alkylation of aniline has been shown to
be catalyzed by finely divided nickel (S1,St, SS,S4,35).
I t was elegantly demonstrated by Rice and Kohn (52)
that alkylation necessarily followed the dehydrogenation of the alcohol. Only under a high enough
temperature, as demonstrated in separate reactions, to
convert alcohol to aldehyde, would alkylation take
place. Nickelalso catalyzes the alkylation of hyd~o-.
carbons (11, IS, 14).
From the standpoint of the alcohol, it is necessary
that the alcohol can be oxidized to an aldehyde. Thus.
tert-butyl alcohol does not alkylate indole-2-carboxylic
acid (21) nor 1,2,3,4-tetraphenylcyclopentadiene (7).
However, triphenylcarbinol does alkylate ethyl malonate (3).
The sum total of the facts cited are consonant with
the following mechanism first stated by Cornforth,
Cornforth, and Robinson (26) for the alkylation of
phenols, by Weizmann, Bergmann, and Sulzbacher for
the Guerbet reaction (28), by Pratt and Frazza for
aniline (Sf), by Sprinzak (8)and by Schoen and Becker
(10) for pentagonal hydrocarbons:

A common structural feature of all of the compounds

which can he alkylated is that they contain an active
methyl, methylene, or a potential active methylene
group. Thus, pyrrole, indole, and phenols fall into
this category because of the potential methylene group
in the structures:

Indeed, it has been shown that 1-methylpyrrole is not

alkylated under the usual reaction conditions (87).
It has also been shown that alcohols with only one
alpha hydrogen fail to undergo the Guerbet reaction
(28). Thus, 2-ethylhexanol gives an ether, but not a
Guerbet product. Compounds with an active hydro-

Equation (a) represents the adventitious oxidation

of the alcohol under the reaction conditions, (b) represents a base-catalyzed condensation of the aldehyde
with the active methylene group, (c) is patterned after
well-known metal-alkoxide interconversion of polarized
bonds, carbonyl groups with alcohols. A more usual
example is that (11). the unsaturated intermediate,
can take part in this reaction because it is a dipolar
molecule (56). Reaction (d) merely completes the
Volume 36, Number 3, March 1959

121

equilibrium, a reasonable step because of the general

greater acidity of alcohols versus hydrocarbons.
Since an unsaturated intermediate is required, this
scheme accounts for the failure t o dialkylate active
methylene groups. It thus accounts for the pure
monoalkylation using this procedure as opposed to the
alkylation with alkyl halides which always gives some
dialkyl product. The first step makes it possible t o
account for the loss of one carbon atom and the monoalkylation with l,3-dials. Here the oxidation would
produce a 8-hydroxyaldehyde, i.e., an aldol. I n the
presence of base the aldol may reverse to give formaldehyde and isobutyraldehyde which could enter into step
( b ) . Searles and Ives (57) have proposed that the
anion of the dial goes t o methanol and an aldehyde
(equations 19, 20). This mechanism appears to be

HOCH,~<

ROH

HOCH.CH<

(20)

credible for cyclic l,3-glycols and for those glycols with

two large substituents at the middle carbon. A third
alternative is the possible direct loss of methanol
(equation 21).

Although direct evidence for the three alternatives is

not available for the general case, the author feels
that the first is the most feasible with some possibility
of the second for highly substituted diols.
An explanation for the failure to produce an alkylated
product with 3-dialkylzmino-l-propanolis not readily
forthcoming. Boiling of the aminoalcohol is accompanied by the evolution of the dialkylamine (18). It is
well known that 3-dialkylaminopropanols decompose to
dialkylamines and or,@-unsaturated aldehydes. Further reaction of the acrolein with base t o form tars
could account, then, for the failure t o obtain the expected alkylated product.
For the origin of 3-benzylquinoline from benzyl
alcohol and quiuoline, Avramoff and Sprinzak (4)
cited the following steps as probable: (1) reduction t o
3,4dihydroquinoline; (2) condensation with benzaldehyde (oxidized benzyl alcohol); (3) further reduction t o
3-benzyl-1,2,3,4-tetrahydroquinoline;or (4) rearrangement t o 3-benzylquiuoline.

122

Journol o f Chemicol Educofion

Literature Cited

HALLER,
A., Compt. tend., 122,1490-94 (1892)
CARROLL,
M. F., J . Chem. Sac., 1940,704-6.
Ibid. 1941, 507-11.
AVIVLMOFF,
M., AND SPRINZAK,
Y., J . Am. Chem. Sac., 78,
4090-96 (1956).

Ibid., 80,493-96 (1958).

SONNTAQ.
N. 0. V.. LINDER.
S.. BECKER.
E.I..

AND

SPOERRI.

LINDER,S. M., BECKER,E. I., AND SPOERRI,P E.,

J.Am.
Chem. Sac., 75,5972-75 (1953).
SPRINZAK,
Y., Bull. Research Council Israel, 3 , 104 (1953).
TSEOU,H.-F., AND CHOWT.-S., J . Chinese Chem. Soe., 7,
27-28 (1939); C. A , , 34, 1985' (1940).
SCHOEN,
K. L., A N D BECKER,E. I., J . Am. Chem. Sac., 77,
603S31 (1955).
N..
1317-18 (1956). '

D.
. . MATTKEWS.

AND

BECKER.E. I.. J . Om. Chem.., 21.,

(12) RURIN,I. I).,A N D BECKER,

E. I., J . Org.Chem., 22,1623-25
(19.57)~
* - - - . ,.
(13) LORENZ,
D. B., B.S. thesis, Polytechnic Ihstitute of Brooklyn, 1958.
(13.1 BENTLEY,
K. W., DOMINGUEZ,
J., AND RINGE,J . P., J . 0ry.
C h . , 22,4W-17 (1957).
(14) HlRsca, S. S., B. S. thesis, Polytechnic Institute of Brooklyn,
1958.
M., Compt. rend., 128,511-3 (1899).
(15) GUERBET,
(16) GUERRET,
M., Bull. soc. chim. France, [3]21, 487-89 (1899).
(17) GUERBET,
M., Bull. 8ac. chim. Franee, [3] 21,489-91 (1899).
(18) MACHEMER,
H., Angew. Chem., 64, 213-20 (1952).
(19) FISCHER,
H., AND BARTEOLOMAUS,
E., Z. Physiol. Chem., 77,
185-201 (1912).
(20)
80. S 1 6 (1912).
, , Ibid..
~,
(21) CORNFORTH,
R. H., AND ROBINBON,
R.,J . Chem. Soe.,
1442. RRII-R2
---(22) Co~ncmcm,U., AND BERTONI,
C., Atti acead. Lincei, 21 1,
653; C.A., 7, 1183 (1913).
(23) PLANCEER,
G., AND ZAMBONINNI,
T.,Atti aeead. Lineei, 21 I ,
598 (1912); C. A,, 7,1183 (1913).
(24)
U..Atti accad. Lincei., 21 I. 489 (1912):
. . COLACICCHI.
., C. A , .
7,1182 (igi3j.
M.,AND SPRINZAK,
Y., J. Ory. Chem., 22, 571(25) AVBAMOFF,
74 (1957).
(26) CORNFORTH,
J. W., CORNFORTH,
R. H., AND ROBINSON,
R.,
J. Chem. Sac., 1942,682-84.
(27) FREIESLEBEN,
W., unpublished work in these laboratories.
(28) WEIZMANN,
Ch., BERGMANN,
E., AND SULZBACHER,
M.,
J . Org. Chem., 15, 54 (1950).
(29) FISCHER,H., AND EISMAYER,
K., Ber., 47, 1820-28 (1914).
(30) SETTEPANI,
J., B.S. thesis, Polytechnic Institute of Brooklyn,

. .

--.

10<4

(31) PRATT,E.F., AND FRAZZA,

E.J., J . Am. Chem. Soe., 76,
6174-79 (1954).
(32) RICE,R. G., AND KOHN,E. J., J . Am. Chem. Soc., 77, 4052
(1955).
,
.
(33) AINSWORTH,
C., J . Am. Chem. Soc., 78,1635 (1956).
(34) CORSON,
B. B., AND DRESSLER,
H., J . Org. Chem., 21, 474
IlOifi!

(35) HORYNA,
J., AND CERN-?,O., Chem. &ty, 50, 381 (1956).
(36) WHELAND,
G. W., AND MANN,D. E., J . Chem. Phys., 17,
264-68 (1949).
(37) SEARLES,
S., JR., AND IVES,
K. E., 127th Meeting of the
American Chemical Society, Cincinnati, Ohio, March 29
to April 7, 1955, Abstracts, p. 24N.